1. Field of the Invention
The invention relates to image processing apparatus and method for processing digital image data and, more particularly, to image processing apparatus and method for performing protection of a copyright, prevention of alteration of an image, a recording of various information, or the like by embedding a digital watermark into digital image data and to a storage medium for realizing such a method.
2. Related Background Art
In recent years, an image is often handled as digital data (digital image data). The digital image data is easily copied by using a computer or the like or easily transmitted through a communication line without deteriorating a picture quality.
Therefore, there is a fear that the digital image data is easily illegally copied and redistributed. A digital watermark technique has been known as a method of suppressing those operations.
The digital watermark technique is mainly classified into: an invisible digital watermark technique for embedding watermark information such as copyright information, user information, or the like in a form such that it is difficult to be identified by the human eyes; and a visible digital watermark technique for forming a watermark image such as a logo or the like of a company which possesses a copyright of an image onto the image in a form such that it can be identified by the human eyes.
When digital watermark information is embedded in an image by the invisible digital watermark technique, whether the watermark information has been embedded or not cannot be identified if the watermark-embedded image is merely seen. Therefore, although it is difficult to delete the watermark information, an illegal copy or an illegal distribution can be easily performed as compared with the visible digital watermark technique.
Even if the digital image data is illegally copied or distributed, since a possibility that the watermark information remains in the digital image data is high, the illegal user can be specified by extracting a user ID or the like embedded as digital watermark information.
To extract the embedded digital watermark information, information showing at which position in the image the digital watermark information has been embedded or associated information (in the case where the embedding position has been determined in a random number manner, an initial value for generating the random numbers, or the like) of such information has to be secretly held by a specific apparatus or person as “key” information for extraction of the digital watermark information.
As representative methods for the invisible digital watermark, there is known a method of performing a frequency conversion such as high speed Fourier transformation, discrete cosine transformation, wavelet transformation, or the like to an input image and embedding a digital watermark into a certain frequency area, a method of directly embedding it into a space area (this method is called a patchwork), or the like. A method of embedding information into a lower bit (for example, least significant bit) showing each pixel value in an image is also known.
Generally, the invisible digital watermark can be used as it is with a value similar to the original image. The information embedded to pursue the illegal use is difficult to be removed from the image. An embedding process of the invisible digital watermark information, therefore, can be regarded as an irreversible image process.
In the case where digital watermark information is embedded in the image by the visible digital watermark technique, the digital watermark information has been embedded onto the digital image in a form such that it is visible by the eyes and it is difficult to use it as it is with a value similar to the original image. Therefore, the act such as illegal copy or illegal distribution can be suppressed.
As a visible type embedding method of the digital watermark information, there is a method whereby each pixel value of an image showing a logo or the like of the copyright owner is exchanged to a value of the m-th bit of a pixel value in a designated area in an original image, thereby embedding copyright information into the original image. Explanation will be made with reference to FIG. 1. Assuming that the least significant bit of the pixel value is the 0-th bit, the m-th bit corresponds to the bit position of a relatively upper bit (bit position near the most significant bit), namely, corresponds to the bit position which can be identified by the human eyes. In FIG. 1, reference numeral 101 denotes an original image; 102 one arbitrary pixel constructing the image; and 103 a bit construction of a pixel value of the pixel 102.
In the case where the user wants to use such an image in which visible digital watermark information has been embedded as an inherent original image, it is necessary to reconstruct the original image by removing the digital watermark information. For this purpose, for example, the value of the m-th bit of the image (original image) before the exchange is secretly held as key information and, when the image is reconstructed to the original image, the value corresponding to the key information is exchanged again to the pixel value of the image in which the digital watermark information has been embedded.
Generally, as for the visible digital watermark information, it is desirable to realize the reversible conversion as mentioned above. If the user intends to accomplish the reversible conversion, however, for example, in the case where the designated area of the image in which visible digital watermark information is embedded occupies most of the original image, there is a problem such that a data amount of the key information to be separately held is very large.
As one of countermeasure methods for the conventional visible digital watermark technique, for example, as shown in FIG. 2, there is a method whereby the key information is embedded as an invisible watermark information into the image and an increase in whole data amount due to the holding of the key information is suppressed. For instance, if the visible watermark key information mentioned above is exchanged to the 0-th bit (the original 0-th bit is deleted), since the 0-th bit is the least significant bit, it is difficult to identify the 0-th bit by the human eyes, the key information can be embedded, and an increase in whole data amount can be also suppressed. If such a method is used, however, the perfect reversible conversion cannot be realized.
Hitherto, if digital watermark information was embedded in an original image by using the invisible digital watermark technique, it is difficult to reconstruct the perfect original image from the image derived after the embedding.
On the other hand, if visible digital watermark information was embedded in an original image, in order to reconstruct the perfect original image from the image derived after the embedding, the key information necessary to remove the digital watermark information is needed and there is a problem that the whole image data amount is increased due to the holding of the key information.